Synthetic identity fraud is different from traditional identity theft in a way that makes it fundamentally harder to detect. An identity thief steals someone's existing identity — their name, their social security number, their credit history — and uses it fraudulently. The victim notices. The bank notices. The fraud is eventually discovered because there is a real person whose records do not match their experience. Synthetic identity fraud creates a person who does not exist, assembled from fragments of real data combined with fabricated elements. There is no victim to notice, because the victim is fictional.

The construction of a synthetic identity typically begins with a real element — often a Social Security number belonging to a child, an elderly person, or a recent immigrant who is unlikely to be actively using their credit file. Around this real anchor, the fraudster builds a fictional person: a fabricated name, a plausible date of birth, an address that passes validation checks. The synthetic identity then begins a process known as "piggybacking" — being added as an authorised user on legitimate accounts to build a credit history that looks genuine.

Over months or years, the synthetic identity develops a credit profile that is indistinguishable from a real person's. It applies for credit cards, takes out loans, makes payments on time, and gradually increases its credit limits. Then, when the credit lines are maximised, the identity "busts out" — maxing out every available credit line and disappearing. The lender is left with losses that cannot be attributed to any real person, and the fraud often is not even classified as fraud in the institution's records. It is written off as a credit loss.

The Federal Reserve has identified synthetic identity fraud as the fastest-growing type of financial crime in the United States, with estimated annual losses exceeding $6 billion. The true figure is likely higher, because the nature of the crime means many losses are never correctly identified as fraud.

Traditional fraud detection systems are designed to identify anomalies in real people's behaviour. They flag transactions that do not match a customer's pattern. They detect when a known fraudulent identity is reused. But they struggle with synthetic identities precisely because these identities have been carefully constructed to look normal. The credit history is clean because it was built to be clean. The spending patterns are unremarkable because they were designed to be unremarkable.

Detection requires a different approach: verifying identity at the point of account creation with sufficient rigour to catch the fabrication before the synthetic identity enters the system. This means going beyond name-and-number matching to include biometric verification that confirms a real, unique human being is behind the application. It means document authentication that detects the increasingly sophisticated forgeries used to support synthetic identities. And it means cross-referencing identity elements against broader data sets to identify combinations that, while individually plausible, are collectively inconsistent.

Generative AI is making synthetic identity construction easier and the resulting identities more convincing. The defensive response must match the sophistication of the attack. deepidv provides identity verification that combines document authentication, biometric matching, and deepfake detection to catch synthetic identities at the point of onboarding — before they enter your system and begin building the credit history that makes them invisible.